Development and characterization of biodegradable microspheres containing selected antimycobacterials
Prolonged therapy required to effectively treat mycobacterial infection frequently results in severe dose-limiting side-effects and drug resistance due to patient non-compliance with protracted dosage regimens. Biodegradable poly-a-hydroxy acid microspheres and microcapsules containing rifampicin (RIF) and isoniazid (INH) respectively have been prepared with the intention of providing high sustained site-specific concentrations to overcome some of the shortcomings of existing oral treatments. Due to the high dose, hydrophilicity and instability of both drugs, formulation strategies to attain high drug loading and methodologies to characterize in vitro drug release during ongoing decomposition were required. Stability indicating HPLC assays to quantify drug release have been developed, validated and applied to monitor drug release based on cumulative quantification of drug and degradates. A mathematical correction for serial decompositions associated with RIF was made based on the terminal pseudo equilibrium observed during stability studies. An isocratic HPLC assay was prospectively developed for the quantification of both drugs and their major metabolites in biological samples. Further preformulation studies confirmed the absence of significant polymorphs for both drugs when recrystallized from solvents later used in formulation development. Furthermore, thermal analysis revealed only modest interaction between the drugs and Resomer®. The high and moderate water solubilities of INH (145 mgmL-1) and RIF (1 mgmL-1) determined the selection of spray-drying (SD) and emulsion solvent evaporation (ESE) for RIF, whereas preparation of INH microcapsules relied solely on the former technique. Examination of the effects of varying RIF: polymer ratio, phase volumes and continuum presaturation with selected poly(L-lactide) and poly(D, L-lactide-co-glycolide) (PDLGA) Resomer® identified optimum conditions to maximize drug loading during a comparison of aqueous ESE with spray-drying with a range of nine further amorphous Resomer® polymers. Although yields were generally higher with ESE (85-90 %), SD (45- 75 %) was considered a superior preparation technique on the basis of the rapid production of microspheres of high and predictable drug loading (100 % of that attempted), with monodisperse granulometry and superior morphology. Release profiles were typically asymptotic characterized by a rapid `burst' of release followed by a slow release of residual entrapped RIF, irrespective of the preparative technique or polymer used. Poor yields (7.2 %) when SD low molecular weight (MW) PDLGA (8 kD) were greatly enhanced (74.8 %) by reduction in drying temperature and substitution of chloroform: dichloromethane (CFM: DCM) (1: 1) cosolvent with DCM. These conditions were adopted as the optimum parameters for further studies of blends of low (2 kD, R104) and moderate (11 kD, R202H) MW poly(D, L-lactide) (PDLLA); materials which demonstrated excellent sprayability and dramatically modulated the release of drug when combined compared to their use alone. Drug release showed a remarkable dependence on blend, dramatic acceleration being observed between 44 and 48 %w/w R104. Release over this range showed a marked dependence on medium temperature and led to the proposal of an autohydration mechanism linked to the hydrophilicity and glass transition (T9) of the blend which accounted for the sigmoidal profiles observed. First order dependence of release allowed calculation of Arrhenius derived activation energies of drug release in glassy anhydrous and rubbery plastic matrices of 630 and 320 J mol-1. Hydration and thermal studies supported the postulated diffusion mechanism, whereas granulometric and morphological examinations demonstrated that erosion did not contribute significantly. The criticality of matrix composition was further highlighted when interchange with nominally identical polymer, R202H, shifted the critical composition to 30 %w/w R104. Moreover, this observation contested the batch-to-batch reproducibility of commercial polymer. Substitution of DCM with halothane (HAL) and acetone (ACT) had a profound influence on the properties of compositionally identical ('R104: R202'H, 30: 70) microspheres, particularly release kinetics. This was attributed to the more rapid drying kinetics with the poor solvent, ACT, and the generation of a porous matrix. Consequently, drug was largely released during the 'burst' phase. Superior solvents, HAL and DCM resulted in enhanced matrix coherence at the expense of considerable residual solvent burdens (6 - 12.5 %), which allowed extensive matrix relaxation as solvent was lost with first order kinetics. This ageing process was followed by the development of an endotherm associated with the Tg as the matrix stabilized with a resultant increase in the induction period and a general retardation of drug release. Extension of the concept of blending R104 as release 'initiator' to a range of MW PDLGA of 50: 50 and 75: 25 comonomer ratio as release 'modulator' was of limited success generating release profiles reminiscent of each polymer when used alone. The magnitude of the 'burst' correlated to the precipitation kinetics of the predominant complementary PDLGA polymer as determined by cloud-point titration. Due to the more hydrophilic nature of the copolymers, at the critical concentration uncontrolled hydration resulted in a single rapid release phase. Spray-dried biodegradable INH microcapsules were prepared by a two stage process whereby SD cores of drug or in combination with biodegradable albumin or casein were subsequently coated with PDLGA by SD. The highly crystalline, aggregated and irregular morphology of SD drug resulted in poor coating efficiency and a rapid release of encapsulated drug. Protein microspheres of superior sphericity allowed more effective coating and hence slower INH release. It is concluded that SD has excellent industrial potential for the preparation of biodegradable poly-a-hydroxy acid microspheres for high dose drugs to be delivered directly to their site of action, e. g., intra-pulmonary. Indeed, the granulometry of these particles and, in particular, the hydrophilic character of blends of PDLLA described have considerable potential for the sustained delivery of drugs in the low volumes of fluid that prevails in the lung. These formulations might offset some of the limitations of current oral antimycobacterial therapy.